Digital video recording technology, such as DVD+RW technology, is rapidly progressing and gaining a large portion of the market for video recordings due to its picture and audio quality and ease of use. One example of the ease of use associated with these video recordings involves navigation of the recordings. Current digital video recording technology divides the video recording into segments called chapters so that users may more easily navigate within the recording, such as by instantly jumping to a particular scene within the video recording.
Typical video recordings include chapters of a fixed length, such as five minutes. However, using fixed length chapters may result in chapter breaks, or boundaries, that occur at inappropriate moments within the video recording. For example, a movie with fixed length chapters may have a chapter boundary right in the middle of an important battle scene or dialogue.
In addition, each chapter generally has a corresponding keyframe in order to identify the chapter for the user. Thus, for example, a user may search through a series of still pictures that are obtained from within the chapters in order to identify the chapter that he or she wishes to view. Currently, these keyframes are selected either automatically or by hand.
However, the currently available methods for automatically selecting keyframes use techniques such as shot break detection. This can result in chapters being divided in the middle of a high cut-rate area, such as may occur in a high action sequence or during a dialogue, as described above. Using a method of selecting keyframes by hand, however, is a cumbersome, personnel intensive process.